Earth Observation Satellite Learning to Video Compressor complexity reduction

• Published in: 2020 IEEE Aerospace Conference pp. 1 - 8
• Main Authors: Bayat, Mohammadreza, Rongke, Liu, Arman, Ladan, Zarrini, Hossein
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.03.2020
• DOI: 10.1109/AERO47225.2020.9172789

Earth Observation (EO) missions in the new era of satellite applications encounter a tangible growth. Capturing live video of the Earth satellites are new trends that would be the next customers' demands for this industry. These rapid changes are having serious effects on studies on customizing this type of mission. The having predictable motion of a satellite in an orbit is a significant advantage for video capturing missions that current study is focusing on. This paper gave a brief overview of some general studies and then discussed the Motion Estimation (ME) technique by means of which Video Compressors are able to detect and eliminate spatial redundancy. The required configuration sections of video capturing satellites in connection with video compression were described as well. To address the main question in this paper which is how to teach video capturing satellites for enhancing video compressor techniques, evidence for the necessity of such complexity reduction was presented. Detecting dominated Motion Vectors (MVs) in statistical aspect for a recorded video sample of an EO satellite put emphasis on the title of this article. The final stage was proposing an algorithm by considering the repetitive manner of EO satellite to scan the earth or any other targets. A simulated scenario in STK software helped us to extend this idea in detailed. Cumulative MVs that stem from the average of the previous MVs repetition round of satellite sweeping, suggest the default MVs for the next round. This procedure could reduce the calculation time that is consumed to find MVs in video compression algorithms. The required memory size based on the defined scenario was calculated.